Keeley DS-1 problems

[MrStab]

hi guys,

well, i thought this would be a fairly straightforward build, but evidently not. i've built the BOSS DS-1 w/ mods as per the layout here: http://www.sabrotone.com/?p=86, which seems true to the schematics here: www.diystompboxes.com/smfforum/index.php?topic=64731.msg511604. granted, i did waste a load of time on a failed JFET buffer, but i've since reverted that back.

so what's the problem? well, there are two problems, and i just don't know where to look for reference voltages etc.
firstly, it distorts even with the Distortion pot at 0, and notes very subtly break up in that "Atari" kinda way as they tail off, which doesn't sound right to me. this is with mod switches off. are either of these characteristic of a Keeley Seeing Eye + Ultra-modded DS-1?

the 2nd problem is reminiscent of DC voltage at the input: crackly guitar pot, general noise. i reduced it from 2 to 0.02V with an extra 10uF input cap and 2.2M resistor to ground. originally, on both the layout and schems there was a 1M from vref going straight to the input with no caps! i'm no expert, but surely that's bad...? but maybe not, if others have successfully built this using those schems.

so i can only assume two things: faulty caps, and/or NPN biasing problem.

iirc, the only parts i've substituted are MC1, which was 47pF and is now 100pF, R9 from Emitter to ground - was 22R and is now 33R, and R7 has gone from 470k to 680k in an attempt to raise input impedance (i'd read others had successfuly done it). there're a coupla instances where i've put resistors in series to get the right value but they work out pretty accurate. oh, and i used a 1uF electrolytic for the output cap instead of a film cap. 0.00V at output.

can't really find voltages or anything on the web - only that my NPN collector voltage (2n5088) is maybe too high at ~5.25V. some folk say a TL072 isn't great for this purpose, but no mention of it being outright problematic. i mean surely i should get a clean-ish sound with distortion on 0, right?

cheers for any help!

Edit: i'll try reverting that 680K back to 470K & see if that changes anything, perhaps i was trying to apply something that worked for others on a stock DS-1 to the modded one.

[Pojo]

I can't immediately help with the issues you're experiencing, but I have an actual Keeley modded DS-1 sitting on the shelf at home. I've never mucked around with the circuit or even studied it at all. But the atari sounds you mention aren't anything I can say describes the sound of it.

If you need me to check any voltages for you just let me know and I'll see what I can do tonight after the kiddies go to bed.

[MrStab]

thanks a lot for the offer, Pojo. voltages of the opamp, NPN buffer & input tip would be really appreciated if you have the time. is there a clean sound when Distortion/Gain is all the way counter-clockwise?

i lowered the 680k to 470k - Collector voltage is now at ~3.9V. things are a bit less noisy, Distortion at 0 is still distorted but a little bit less so. i had a vaguely similar issue ages ago which was resolved by using shielded in/out wires, but that was using a layout i'd made myself, which was flawed. the notes still break up abnormally at the end.

just remembered another substitution (whilst trying to figure out why dist. has that floor): at the 2.4k and 1uF cap to ground (4.7k and 470uF stock), i put 3.3K.

[Pojo]

is there a clean sound when Distortion/Gain is all the way counter-clockwise?

Its been a while since I've played through it and I don't remember off hand how clean it got at counter clockwise....I'll be sure to check that as well. I take it you're not getting much variation from minimum to max distortion?

[MrStab]

there is variation throughout the pot's turn, but there's just no clean.

i've changed pretty much all my substitutes to the values on the BOM to no avail (apart from using a TL072).

i thought maybe the gain could be amplifying scratchyness in my guitar pot, but no other high-gain circuits have done this in the past. currently measuring 0.06V at input, surely that's too low to be the problem? i'm sure i've read "to within a few milivolts" regarding what's okay and what's not here. gah. distortions are meant to be simple!!!! lol

FWIW, i've tried 3 x 2n5088's throughout all this as well, as well as a fresh opamp

[MrStab]

2:08 in this vid:

Distortion full CCW, yet no clean. so maybe that clears that up. if so, the scratchy guitar pot and note break-up remain issues. another description for the note break-up would be "torn-speaker-like".

i could try to get a sample tomorrow, but my camera mic wouldn't be much help and i recently reinstalled Windows so no installed recording software unfortunately. except that horrific open-source one i wouldn't urinate on to put out a fire.

[MrStab]

tried a shielded input wire, no improvement. running out of options. i'm only gonna give this another day or so as i don't feel it's worth the hassle for what should be a fairly-straightforward circuit.

i was originally gonna build an OCD, but i didn't like the sound of its impedance issues - i just want a good distortion or OD with minimal end-user annoyance, as i plan to give it away (to no-one in particular). might just have to make another TS-808.

[Pojo]

Well before you give up completely, here's some voltages as promised. Something to compare to, right?

Supply: 9.6V

M5223AL SIP dual opamp:
1: 4.78V
2: 4.78V
3: 4.72V
4: 0V
5: 4.78V
6: 4.78V
7: 4.78V
8: 9.57V

Q1 (2SC2240 NPN):
E: 3.64V
C: 9.57V
B: 4.06V

Q2 (2SC2240 NPN):
E: 12.1mV
C: 4.29V
B: .601V

Q3 (2SC2240 NPN):
E: 3.16V
C: 9.57V
B: 3.47V

Q4 (2SC2240 NPN):
E: 0
C: 93.7mV
B: .674V

Q5 (2SC2240 NPN):
E: 0
C: 24.4mV
B: .620V

Q6 (2SK30A JFET N channel):
S: 4.74V
G: 0
D: 4.74V

Q7 (2SK30A JFET N channel):
S: 4.74V
G: fluctuates wildly
D: 4.76V

Q8 (2SK30A JFET N channel):
S: 4.74
G: 57mV (unstable, would start at that reading then slowly go down)
D: 4.76

[MrStab]

thanks loads, Pojo! i really appreciate you going to all that trouble, and i owe it to you not to give up now!!

i won't need the JFET voltages, but the rest should be really useful. stupidly, i didn't think that the opamp would have "standard" vbias voltages as despite looking at the schematic a lot, and the configuration being clear as day, for some reason i thought it was being used in some less-typical configuration. from what i remember, at least i seem to have Q3's collector right. i won't have time to work on it till sunday or monday but i will as soon as i can.

i still don't understand how, in the modded schematics, there's that 1M resistor from the vbias without a cap to prevent it going back to the input. probably some explanation i still need to learn.

at around 1am yesterday, i started work on a Tube Screamer to restore some confidence, somehow i finished it quite quickly lol:



i stole the 20K pot from the DS-1 but i'll sub it with a 10k or a 50k temporarily when i get back to it.

cheers!

[Pojo]

Happy to help and nice one on the TS!

I'm actually only now taking a look at the schematic. You raise a good point about the 1M resistor and I think they just chose vbias as an alternative to ground...but it looks to be put in as a pull down resistor ironically enough. I can't see how DC would not make it's way out through the input. It wouldn't be biasing the NPN because of C3 and since R6, R7, and R8 are taking care of that. I'd try R5 to ground and see if that cures the scratching.

Also maybe go through the circuit with a signal probe to narrow down where the odd breakup is coming from.

[duck_arse]

it's a bit late now, but I would say the 1M resistor at the input to Vref is just a drawing error, unfortunately carried through to the pcb layout. it should go to ground, as you say.

[MrStab]

i'm getting quite fond of Ye Olde Probe (i wired a cap to a jack so i can use any guitar lead) so it was my first port of call when my JFET buffer failed, but *iirc* the unusual breakup was hard to locate for whatever reason - i'll look at it again.

so that 1M is just a standard pulldown resistor? hmm... maybe everyone's been using this with other circuits in front so they haven't noticed issues with it, or maybe i just got unlucky.

when i get back to work i'll send that 1M to ground, then. i've been trying to suffer lead-free solder with this one for legality's sake, but it's making alterations a nightmare. thanks w/ the TS, it was fun making a second one - my first has become a staple in my live rig.

cheers guys, i'll report back

[MrStab]

ok, so cutting that 1M resistor (and relying on a pre-cap 2.2M resistor) and replacing the opamp has fixed the guitar pot noise and now nothing sound's straight-up wrong with the distortion per se, but it feels like there's not the right amount of sustain you'd expect. like there's some loading i can't seem to find. i can't see it being the 2.2M as it had no notable effect on the signal when i'd removed it previously and i figure pulldowns coming before the cap is fairly standard.

[MrStab]

it's seeming quite likely now that everything works fine, but taste is clouding my judgement. too much low end! i'm notorious for my treble (maybe i need to over-compensate for hearing loss). it'd help to actually have/played a Keeley DS-1 to compare to, but part of the reason i do DIY is because i'm broke! lol

so nothing "wrong" with  it per se - it just seems impossible to dial in a strong high end. maybe lowering the input cap from 100nF to the stock 47nF would be of some benefit, but as it's not for me, and a lot of co-musicians i know are into sub-bass ultra black death metal, i might just leave it as is.

i've done a lot of trawling round the web and have inferred that it's normal for sustain to go down a little bit compared to stock (for some) and the Gain pot's taper does get screwed up - so i might put a reverse log in there to even things out.

any thoughts?

[duck_arse]

R13//C8, under the gain pot, set the low-end gain of the ic1b, it's a little hard to read, if you are using page 8 of the millar/iqbal docs. reduce the cap  will raise the bass cutoff. the resistor will change the gain.

[Pojo]

Hmm, I think something must be different in your build. In the Keeley DS1 that I have there are plenty of highs, in fact I needed to keep the tone at around 9 o'clock to keep them tamed IIRC. Could just be the tolerance of the parts used or something like that.

[MrStab]

hmm... i'm tempted to change R13/C8 as you suggested, Duck, but if Pojo's DS-1 is more trebly, that makes me think there's something up with my build. can't see any obvious wrong values, as it stands.

as i had to change the opamp (i'm always killing them), it follows that maybe i should try changing Q2 as well, though ime a dead NPN just produces no signal. i'll have another look and switch R13 & C8 back to stock if nothing jumps out at me.

some online retailer just sent me a really long parcel just so they didn't have to split a bunch of IC sockets into two tubes! thought i'd mention that here as no-one else would know what i'm on about. tsk!

cheers!